State monitoring method and monitoring system for photovoltaic panel waste heat coupling power generation equipment

Abstract

The invention discloses a state monitoring method and monitoring system for photovoltaic panel waste heat coupling power generation equipment. The method comprises the following steps: acquiring operation original data; classifying and denoising the original data; cleaning and fusing the de-noised data to obtain standardized data; establishing a health index model based on the standardized data, calculating a health index and comparing the health index with a threshold value; when the health index is abnormal, data dimension reduction is carried out, and the fault type and position are determined; system parameters are adjusted or graded alarming is performed according to fault types; through multi-sensor data fusion and dynamic health index evaluation, and in combination with multi-stage fault diagnosis of principal component analysis dimensionality reduction, support vector machine classification and isolated forest detection, the crossing from single parameter alarm to system state comprehensive diagnosis is realized, the fault identification accuracy is improved, and the fault diagnosis efficiency is improved. And the problems of high false alarm rate and difficulty in positioning multiple faults caused by a static threshold and a single algorithm in the prior art are solved, and the operation reliability and the energy recovery efficiency of the waste heat coupling power generation system are remarkably improved.

Description

Technical Field

[0001] The present invention relates to a photovoltaic panel waste heat coupled power generation system, and in particular to a state monitoring method and a monitoring system for photovoltaic panel waste heat coupled power generation equipment. Background Art

[0002] With the widespread adoption of solar photovoltaic technology, photovoltaic panels generate significant heat during operation, which not only reduces their photoelectric conversion efficiency but also potentially shortens their service life. Recycling waste heat from photovoltaic panels can further improve energy efficiency, reduce energy waste, lower carbon emissions, and achieve sustainable energy utilization. This not only helps improve the economic viability of renewable energy, but also provides strong support for energy conservation, emission reduction, and environmental protection, promoting the development and application of green energy technologies.

[0003] The applicant previously applied for patent application number 202510336925.1 for a photovoltaic cooling and waste heat utilization device. The main content is "The photovoltaic cooling and waste heat utilization device includes a photovoltaic panel, a heat storage tank, a flat heat pipe and a carrier. The heat storage tank is composed of a box body, thermal conductive ceramics, a thermoelectric power generation sheet, a phase change material, etc., which can effectively recover the waste heat generated by the photovoltaic panel during operation and convert it into electrical energy through the thermoelectric power generation sheet. At the same time, the phase change material is used to cool the photovoltaic panel to improve the power generation efficiency. The flat heat pipe transfers the waste heat to external equipment for further utilization, achieving diversified waste heat recovery and conversion. The flat heat pipe is also used to reversely input heat into the heat storage tank when the temperature of the photovoltaic panel is lower than the set temperature. The device has a compact structure, good thermal conductivity and heat preservation, can maintain stable operation in different environments, and improve the overall performance of the photovoltaic system."

[0004] These photovoltaic panel waste heat-coupled power generation systems are complex and operate in a volatile environment. Condition monitoring is essential to monitor operating parameters in real time, detect potential faults, optimize efficiency, and extend equipment life, thereby improving system reliability and cost-effectiveness. Existing technologies typically use single sensors to monitor key parameters, such as heat pipe temperature or power generation unit power, and rely on static thresholds or simple rule engines for fault diagnosis. However, these approaches have significant limitations. Due to the lack of multi-sensor data fusion and dynamic health assessment, the system struggles to distinguish measurement noise from true fault signals, resulting in a high false alarm rate. Furthermore, fixed thresholds cannot adapt to equipment aging or fluctuating operating conditions, making it difficult to detect gradual faults. Furthermore, when multiple faults occur simultaneously, existing technologies often only trigger general alarms and are unable to precisely locate the source of the fault, forcing maintenance personnel to shut down the system for investigation. This not only increases unplanned downtime but also reduces energy recovery efficiency, severely limiting the reliability and cost-effectiveness of waste heat power generation systems. Summary of the Invention

[0005] Purpose of the invention: The purpose of the present invention is to provide a condition monitoring method for photovoltaic panel waste heat coupled power generation equipment that can realize multi-parameter collaborative monitoring, dynamic health assessment and intelligent fault location. Another purpose of the present invention is to provide a condition monitoring system for photovoltaic panel waste heat coupled power generation equipment.

[0006] Technical solution: The method for monitoring the status of a photovoltaic panel waste heat coupled power generation device according to the present invention comprises the following steps:

[0007] S1: The raw data of heat pipe temperature, water tank temperature gradient, thermoelectric power generation unit power, photovoltaic panel temperature and thermal storage tank phase change material temperature parameters are collected in real time through a multi-sensor network;

[0008] S2: Classify and denoise the collected raw data to obtain denoised operating data;

[0009] S3: Detect and clean the denoised operating data, process duplicate values, missing values, outliers, and erroneous data, and then fuse the same parameters collected by multiple sensors after cleaning to obtain standardized data;

[0010] S4: A five-dimensional health index model is established based on standardized data. The health indexes of the heat pipe, water tank, thermoelectric power generation unit, photovoltaic panel, and thermal storage tank are calculated and compared with preset thresholds. The model achieves a comprehensive assessment of the system status through the multi-dimensional coupling relationship of each health index.

[0011] S5: When the health index of any dimension exceeds the threshold, the data is reduced in dimension through principal component analysis, the fault type is classified using support vector machine, and the abnormal data cluster is identified through isolation forest model to determine the specific location and type of the fault;

[0012] S6: Based on the fault type, dynamically adjust system parameters by combining PID control, predictive function control, and linear quadratic control, or trigger a graded alarm mechanism.

[0013] Preferably, the denoising process in step S2 includes:

[0014] Based on the collected raw data, the LMS adaptive filtering algorithm is used based on the photovoltaic panel temperature data, heat pipe temperature data and phase change material temperature data of the heat storage tank of the waste heat coupled power generation equipment to dynamically adjust the filter parameters according to the data characteristics to remove high-frequency noise;

[0015] Based on the water tank temperature gradient data and thermoelectric power generation unit power data, which are easily affected by environmental interference, the wavelet transform algorithm is used, combined with the soft threshold and hard threshold methods, to perform multi-scale decomposition and reconstruction of the data to further remove noise interference in the data.

[0016] Preferably, the data cleaning in step S3 is specifically as follows:

[0017] Based on the various working parameters of the completed denoising process, the isolation forest algorithm is used to detect the denoised operating data; for duplicate values, a unique identifier is set to identify and delete duplicate records to ensure the uniqueness of the data, and a hash algorithm is used to improve the comparison speed and deduplication efficiency; for missing values, when the proportion of missing values ​​is high and the impact on the analysis is small, the records containing missing values ​​are directly deleted; for time series data, a linear interpolation method can be used to fill missing values ​​with the mean, median or mode; or missing values ​​can be marked as special values ​​for special processing in subsequent analysis; for erroneous data, if the format is wrong, the format is corrected; if the logic is wrong, the logic error is corrected through logical judgment, and the values ​​outside the reasonable range are corrected; for outliers, when the outliers have a large impact on the analysis results, they are directly deleted; when the outliers have a small impact on the analysis results, the outliers are replaced with values ​​within a reasonable range to obtain the data after preliminary cleaning;

[0018] A dynamic time warping algorithm is used to calculate the characteristic similarity between the normal operating condition fluctuations and equipment abnormalities of the data after preliminary cleaning, and the photovoltaic panel temperature data, thermoelectric power generation unit power data, water tank temperature data, heat pipe temperature data, and thermal storage tank phase change material temperature data after precise cleaning are obtained;

[0019] Based on the photovoltaic panel temperature data, heat pipe temperature data, heat storage tank phase change material temperature data, water tank temperature gradient data, and thermoelectric power generation unit power data after data cleaning, the photovoltaic panel temperature data and thermoelectric power generation unit power data are collected by a single sensor and are directly used to calculate the health index after data cleaning; the heat pipe temperature data, heat storage tank phase change material temperature data, and water tank temperature gradient data are the same parameters collected by multiple sensors, and the Kalman filter algorithm is used for data fusion to improve the accuracy and reliability of the data.

[0020] Preferably, the health index of the heat pipe, water tank, thermoelectric power generation unit, photovoltaic panel and heat storage tank in step S4 is calculated as follows:

[0021] The heat pipe health index is calculated based on the temperature difference between the two ends of the heat pipe and the thermal conductivity efficiency. The formula is:

[0022]

[0023] T 冷端 and T 热端 are the temperatures of the cold and hot ends of the heat pipe, T 设定值 is the preset temperature difference;

[0024] The water tank health index is calculated based on the temperature gradient and water flow rate in different areas of the water tank. The formula is:

[0025]

[0026] T i is the temperature of different areas in the water tank, T 平均 is the average temperature of the water tank, N is the number of sampling points;

[0027] The health index of the thermoelectric power generation unit is calculated based on the power generation power and temperature difference. The formula is:

[0028]

[0029] P 实际 is the actual generated power, P 理论 is the theoretical power generation power;

[0030] The photovoltaic panel health index is calculated based on the photovoltaic panel temperature and power generation power. The formula is:

[0031]

[0032] P 实际 and P 理论 are the actual power generation and theoretical power generation respectively, T 设定值 and T 实际 They are set temperature and actual temperature respectively;

[0033] The heat storage tank health index is calculated based on the phase change material temperature and the temperature difference between the inside and outside. The formula is:

[0034]

[0035] T 内部 and T 内部 are the internal and external temperatures of the thermal storage tank, respectively.

[0036] Preferably, step S5 specifically includes: setting a preset threshold based on the evaluation result of the health index on the operating status of each component, and triggering the fault detection mechanism when the HI index exceeds the preset threshold;

[0037] The fault detection mechanism is a data-driven method. The establishment process is as follows: based on the data characteristics of the system operation data after data processing, it is judged whether it meets the preset threshold of the data. If it does not meet the standard, the principal component analysis algorithm is used to reduce the dimension. If it meets the standard, the support vector machine method is used to classify and predict faults for each data item.

[0038] At the same time, based on the data characteristics after dimensionality reduction, the random forest algorithm is used to perform regression analysis on the data. If the regression analysis results show that the data accuracy reaches the threshold, the final processing results are output;

[0039] Based on the final output result, an isolation forest model is constructed and trained to obtain a trained isolation forest model. The model is used to identify abnormal data and trends and determine the specific location and type of the fault.

[0040] Preferably, the adaptive adjustment in step S6 includes: dynamically adjusting the temperature difference at both ends of the heat pipe through a PID controller to adjust the heat pipe blockage fault problem; optimizing the cooling system response speed through a predictive function controller to adjust the photovoltaic panel temperature abnormality problem; and balancing the heat absorption and release rate of the phase change material through a linear quadratic regulator to adjust the heat storage tank temperature abnormality problem.

[0041] A condition monitoring system for photovoltaic panel waste heat coupled power generation equipment of the present invention comprises:

[0042] Data acquisition module, used to collect real-time data on heat pipe temperature, water tank temperature gradient, thermoelectric power generation unit power, photovoltaic panel temperature, and phase change material temperature of the thermal storage tank;

[0043] The data processing module includes a denoising module, a data cleaning module, and a data fusion module, where:

[0044] The denoising processing module uses an adaptive filtering algorithm to denoise the photovoltaic panel temperature, heat pipe temperature and heat storage tank phase change material temperature, and uses a wavelet transform algorithm to denoise the water tank temperature gradient and the thermoelectric power generation unit power;

[0045] The data cleaning module processes outliers and missing values ​​using the dynamic time warping algorithm and the isolation forest algorithm;

[0046] The data fusion module uses the Kalman filter algorithm to fuse the same parameters collected by multiple sensors;

[0047] The condition monitoring module defines the health index of heat pipes, water tanks, thermoelectric power generation units, photovoltaic panels, and thermal storage tanks based on data processing results, and evaluates the operating status of each component using preset thresholds;

[0048] Fault diagnosis module, used to classify and locate faults based on the isolation forest model and support vector machine when the health index exceeds the threshold;

[0049] The system control module is used to trigger parameter adjustment or alarm mechanism according to the fault diagnosis results.

[0050] Preferably, the data acquisition module includes: multiple groups of temperature sensors arranged at both ends of the heat pipe; a temperature gradient monitoring network distributed in different areas of the water tank; a K-type thermocouple module for collecting the voltage and current of the temperature difference power generation unit; a temperature sensor arranged on the bottom surface of the photovoltaic panel; and an embedded sensor for monitoring the temperature difference between the inside and outside of the heat storage tank.

[0051] Preferably, the adaptive filtering algorithm in the denoising processing module is the LMS algorithm, which dynamically adjusts the filter order to match the noise spectrum; a wavelet transform algorithm is used, combined with soft threshold and hard threshold methods to perform multi-scale decomposition and reconstruction to remove noise interference in the data.

[0052] Preferably, the fault diagnosis module includes: using a principal component analysis algorithm to reduce the dimension of the data; performing regression analysis through a random forest algorithm to verify data accuracy; and constructing an isolation forest model to identify abnormal trends.

[0053] Preferably, the system control module includes a PID controller for adjusting the temperature difference of the heat pipe, a predictive function controller for optimizing the response speed of the photovoltaic panel cooling system, and a linear quadratic regulator for balancing the heat absorption and release process of the phase change material of the heat storage tank.

[0054] Beneficial effects: Compared with the existing technology, the present invention has the following significant advantages: through multi-sensor data fusion and dynamic health index evaluation, combined with principal component analysis dimensionality reduction, support vector machine classification and isolation forest detection of the multi-level fault diagnosis architecture, a complete set of intelligent monitoring and control systems is formed, which realizes the leap from single parameter alarm to comprehensive diagnosis of system status, improves the fault identification accuracy, and solves the problems of high false alarm rate and difficult location of multiple faults caused by static thresholds and single algorithms in the existing technology, significantly improving the operation reliability and energy recovery efficiency of the waste heat coupled power generation system. BRIEF DESCRIPTION OF THE DRAWINGS

[0055] Figure 1 This is a flow chart of the method for monitoring the status of photovoltaic panel waste heat coupled power generation equipment of the present invention.

[0056] Figure 2 Schematic diagram of the system structure of the present invention. DETAILED DESCRIPTION

[0057] The technical solution of the present invention will be further described below with reference to the accompanying drawings.

[0058] like Figure 1 As shown, this embodiment provides a method and system for monitoring the status of a photovoltaic panel waste heat coupled power generation device, the method comprising the following steps:

[0059] S1: The original data of heat pipe temperature, water tank temperature gradient, thermoelectric power generation unit power, photovoltaic panel temperature and thermal storage tank phase change material temperature parameters are collected in real time through a multi-sensor network.

[0060] In a photovoltaic panel waste heat coupled power generation system, data acquisition is the foundation of the entire monitoring system. Temperature sensors are placed at both ends of the heat pipe, multiple sets of temperature sensors are placed in different areas of the water tank, K-type thermocouples are used to collect voltage and current data from the thermoelectric power generation module, temperature sensors are placed on the bottom surface of the photovoltaic panel, and embedded temperature sensors monitor the temperature difference between the inside and outside of the thermal storage tank and the temperature changes of the phase change material. Raw data of key parameters is collected in real time at a set collection frequency. Data collected by these sensors is transmitted to the central control system via WiFi, ensuring real-time and accurate data.

[0061] Specifically, the comprehensiveness and accuracy of data collection directly determine the reliability of subsequent status monitoring and fault diagnosis. Accurately collecting heat pipe temperature provides a timely reflection of the heat pipe's thermal conductivity, while monitoring the water tank's temperature gradient helps assess the uniformity of heat distribution within the tank. By deploying sensors at multiple locations, a temperature gradient monitoring network can be formed, providing a comprehensive understanding of the system's internal thermodynamic state. In practice, temperature sensors placed at both ends of the heat pipe monitor its temperature distribution in real time (this example is used for, but is not limited to, heat pipes). These sensors record temperature data at both the cold and hot ends of the heat pipe at a preset frequency (e.g., once per second). This data is transmitted in real time to a central control system. By analyzing the temperature difference, the system can initially determine whether the heat pipe's thermal conductivity is normal. A sudden increase in the temperature difference may indicate a blockage risk, requiring further diagnosis. During data transmission, the system uses a WiFi module to ensure real-time data transmission. A data storage path is also established, storing the collected data in a local database for subsequent analysis.

[0062] S2: Classify and denoise the collected raw data to obtain denoised operating data.

[0063] Based on the collected raw data, the LMS adaptive filtering algorithm is used based on the photovoltaic panel temperature data, heat pipe temperature data and phase change material temperature data of the heat storage tank of the waste heat coupled power generation equipment to dynamically adjust the filter parameters according to the data characteristics to remove high-frequency noise;

[0064] Based on the water tank temperature gradient data and thermoelectric power generation unit power data, which are easily affected by environmental interference, the wavelet transform algorithm is used, combined with the soft threshold and hard threshold methods, to perform multi-scale decomposition and reconstruction of the data to further remove noise interference in the data.

[0065] S3: Detect and clean the denoised operating data, process duplicate values, missing values, outliers, and erroneous data, and then fuse the same parameters collected by multiple sensors after cleaning to obtain standardized data;

[0066] Based on the various working parameters of the completed denoising process, in order to ensure the accuracy and consistency of the data, the isolation forest algorithm is used to detect the denoised operating data; for duplicate values, a unique identifier is set to identify and delete duplicate records to ensure the uniqueness of the data, and a hash algorithm is used to improve the comparison speed and deduplication efficiency; for missing values, when the proportion of missing values ​​is high and the impact on the analysis is small, the records containing missing values ​​are directly deleted; for time series data, linear interpolation methods can be used to fill missing values ​​with mean, median or mode; or missing values ​​can be marked as special values ​​for special processing in subsequent analysis; for erroneous data, if the format is wrong, the format is corrected; if the logic is wrong, the logic error is corrected through logical judgment, and the values ​​outside the reasonable range are corrected; for outliers, when the outliers have a large impact on the analysis results, they are directly deleted; when the outliers have a small impact on the analysis results, the outliers are replaced with values ​​within a reasonable range to obtain the data after preliminary cleaning;

[0067] A dynamic time warping algorithm is used to calculate the characteristic similarity between the normal operating condition fluctuations and equipment abnormalities of the data after preliminary cleaning, and the photovoltaic panel temperature data, thermoelectric power generation unit power data, water tank temperature data, heat pipe temperature data, and thermal storage tank phase change material temperature data after precise cleaning are obtained;

[0068] Based on the photovoltaic panel temperature data, heat pipe temperature data, heat storage tank phase change material temperature data, water tank temperature gradient data, and thermoelectric power generation unit power data after data cleaning, the photovoltaic panel temperature data and thermoelectric power generation unit power data are collected by a single sensor and are directly used to calculate the health index after data cleaning; the heat pipe temperature data, heat storage tank phase change material temperature data, and water tank temperature gradient data are the same parameters collected by multiple sensors, and the Kalman filter algorithm is used for data fusion to improve the accuracy and reliability of the data.

[0069] Specifically, the Kalman filter's state vector and covariance matrix are initialized. Then, at each time step, the current state is estimated based on the previous state and the measured value is predicted. Finally, an update step is performed to correct the predicted value based on the current measured value, resulting in a more accurate state estimate. This dynamic adjustment allows the system to integrate data from multiple sensors and obtain more accurate temperature values. This not only improves data accuracy but also enhances system robustness, making it more resilient to sensor failures or data loss.

[0070] S4: A five-dimensional health index model is established based on standardized data. The health indexes of the heat pipe, water tank, thermoelectric power generation unit, photovoltaic panel, and heat storage tank are calculated respectively and compared with the preset thresholds. The model realizes a comprehensive assessment of the system status through the multi-dimensional coupling relationship of each health index.

[0071] The health index is a core indicator for assessing system health. By quantifying the performance of each component, potential problems can be identified promptly. Thresholds are set based on historical data and expert experience to ensure the reliability of the assessment results.

[0072] The health index of heat pipes, water tanks, thermoelectric power generation units, photovoltaic panels and thermal storage tanks is calculated as follows:

[0073] The heat pipe health index is calculated based on the temperature difference between the two ends of the heat pipe and the thermal conductivity efficiency. The formula is:

[0074]

[0075] T 冷端 and T 热端 are the temperatures of the cold and hot ends of the heat pipe, T 设定值 is the preset temperature difference;

[0076] The water tank health index is calculated based on the temperature gradient and water flow rate in different areas of the water tank. The formula is:

[0077]

[0078] T i is the temperature of different areas in the water tank, T 平均 is the average temperature of the water tank, N is the number of sampling points;

[0079] The health index of the thermoelectric power generation unit is calculated based on the power generation power and temperature difference. The formula is:

[0080]

[0081] P 实际 is the actual generated power, P 理论 is the theoretical power generation power;

[0082] The photovoltaic panel health index is calculated based on the photovoltaic panel temperature and power generation power. The formula is:

[0083]

[0084] P 实际 and P 理论 are the actual power generation and theoretical power generation respectively, T 设定值 and T 实际 They are set temperature and actual temperature respectively;

[0085] The heat storage tank health index is calculated based on the phase change material temperature and the temperature difference between the inside and outside. The formula is:

[0086]

[0087] T 内部and T 内部 are the internal and external temperatures of the thermal storage tank, respectively.

[0088] S5: When the health index of any dimension exceeds the threshold, the data is reduced in dimension through principal component analysis, the fault type is classified using support vector machine, and the abnormal data cluster is identified through isolation forest model to determine the specific location and type of the fault.

[0089] Based on the evaluation results of the health index on the operating status of each component, a preset threshold is set. When the HI index exceeds the preset threshold, the fault detection mechanism is triggered;

[0090] The fault detection mechanism is a data-driven method. The establishment process is as follows: based on the data characteristics of the system operation data after data processing, it is judged whether it meets the preset threshold of the data. If it does not meet the standard, the principal component analysis algorithm is used to reduce the dimension. If it meets the standard, the support vector machine method is used to classify and predict faults for each data item.

[0091] At the same time, based on the data characteristics after dimensionality reduction, the random forest algorithm is used to perform regression analysis on the data. If the regression analysis results show that the data accuracy reaches the threshold, the final processing results are output;

[0092] Based on the final output result, an isolation forest model is constructed and trained to obtain a trained isolation forest model. The model is used to identify abnormal data and trends and determine the specific location and type of the fault.

[0093] S6: Based on the fault type, dynamically adjust system parameters by combining PID control, predictive function control, and linear quadratic control, or trigger a graded alarm mechanism.

[0094] Fault handling and early warning ensure the safety and reliability of the system. By automatically adjusting system parameters, the system can respond quickly when a fault occurs, reducing downtime. The alarm mechanism reminds maintenance personnel to promptly address the fault and avoid potential safety risks. Taking an abnormal water tank temperature gradient as an example, if an abnormal water tank temperature gradient is detected, the system will automatically adjust the valve opening and closing status, optimize the water flow path and the water tank temperature gradient, and ensure the normal operation of the system. During specific execution, the system selects the appropriate control strategy based on the type and severity of the fault. The multivariable and multi-model adaptive decoupling control strategy ensures stable operation of the system during fault handling by dynamically adjusting control parameters. At the same time, the system triggers the alarm mechanism, alerting maintenance personnel to conduct inspections and repairs through audible and visual alarms or remote notifications. The PID control algorithm adjusts the control variable in real time through a combination of proportional, integral, and differential terms to ensure that system parameters remain stable within the set range.

[0095] like Figure 2As shown, this embodiment also provides a condition monitoring system for photovoltaic panel waste heat coupled power generation equipment, including:

[0096] The data acquisition module is used to collect the real-time temperature of the heat pipe, the temperature gradient of the water tank, the power of the thermoelectric power generation unit, the temperature of the photovoltaic panel, and the temperature of the phase change material of the heat storage tank; multiple sets of temperature sensors are arranged at both ends of the heat pipe; a temperature gradient monitoring network is distributed in different areas of the water tank; a K-type thermocouple module is used to collect the voltage and current of the thermoelectric power generation unit; a temperature sensor is installed on the bottom surface of the photovoltaic panel; and an embedded sensor is used to monitor the temperature difference between the inside and outside of the heat storage tank.

[0097] The data processing module includes a denoising module, a data cleaning module, and a data fusion module, where:

[0098] The denoising processing module uses an adaptive filtering algorithm to denoise the photovoltaic panel temperature, heat pipe temperature, and phase change material temperature of the heat storage tank. It also uses a wavelet transform algorithm combined with soft threshold and hard threshold methods to perform multi-scale decomposition and reconstruction to denoise the water tank temperature gradient and the power of the thermoelectric power generation unit. The adaptive filtering algorithm is the LMS algorithm, which dynamically adjusts the filter order to match the noise spectrum.

[0099] The data cleaning module processes outliers and missing values ​​through the dynamic time warping algorithm and the isolation forest algorithm.

[0100] The data fusion module uses the Kalman filter algorithm to fuse the same parameters collected by multiple sensors.

[0101] The condition monitoring module defines the health index of heat pipes, water tanks, thermoelectric power generation units, photovoltaic panels and heat storage tanks based on data processing results, and evaluates the operating status of each component through preset thresholds.

[0102] The fault diagnosis module is used to classify and locate faults based on the isolation forest model and support vector machine when the health index exceeds the threshold; use the principal component analysis algorithm to reduce the dimension of the data; perform regression analysis through the random forest algorithm to verify data accuracy; and construct an isolation forest model to identify abnormal trends.

[0103] The system control module is used to trigger parameter adjustments or alarm mechanisms based on fault diagnosis results; it includes a PID controller for adjusting the temperature difference of the heat pipe, a predictive function controller for optimizing the response speed of the photovoltaic panel cooling system, and a linear quadratic regulator for balancing the heat absorption and release process of the phase change material in the thermal storage tank.

Claims

1. A method for monitoring the status of a photovoltaic panel waste heat coupled power generation device, characterized in that: The following steps are involved: S1: The raw data of heat pipe temperature, water tank temperature gradient, thermoelectric power generation unit power, photovoltaic panel temperature and thermal storage tank phase change material temperature parameters are collected in real time through a multi-sensor network; S2: Classify and denoise the collected raw data to obtain denoised operating data; S3: Detect and clean the denoised operating data, process duplicate values, missing values, outliers, and erroneous data, and then fuse the same parameters collected by multiple sensors after cleaning to obtain standardized data; S4: A five-dimensional health index model is established based on standardized data. The health indexes of the heat pipe, water tank, thermoelectric power generation unit, photovoltaic panel, and thermal storage tank are calculated and compared with preset thresholds. The model achieves a comprehensive assessment of the system status through the multi-dimensional coupling relationship of each health index. S5: When the health index of any dimension exceeds the threshold, the data is reduced in dimension through principal component analysis, the fault type is classified using support vector machine, and the abnormal data cluster is identified through isolation forest model to determine the specific location and type of the fault; S6: Based on the fault type, dynamically adjust system parameters by combining PID control, predictive function control, and linear quadratic control, or trigger a graded alarm mechanism.

2. The condition monitoring method according to claim 1, characterized in that: The denoising process in step S2 includes: Based on the collected raw data, the LMS adaptive filtering algorithm is used based on the photovoltaic panel temperature data, heat pipe temperature data and phase change material temperature data of the heat storage tank of the waste heat coupled power generation equipment to dynamically adjust the filter parameters according to the data characteristics to remove high-frequency noise; Based on the water tank temperature gradient data and thermoelectric power generation unit power data, which are easily affected by environmental interference, the wavelet transform algorithm is used, combined with the soft threshold and hard threshold methods, to perform multi-scale decomposition and reconstruction of the data to further remove noise interference in the data.

3. The condition monitoring method according to claim 1, characterized in that: The data cleaning in step S3 is specifically as follows: Based on the various working parameters of the completed denoising process, the isolation forest algorithm is used to detect the denoised operating data; for duplicate values, a unique identifier is set to identify and delete duplicate records to ensure the uniqueness of the data, and a hash algorithm is used to improve the comparison speed and deduplication efficiency; for missing values, when the proportion of missing values ​​is high and the impact on the analysis is small, the records containing missing values ​​are directly deleted; for time series data, a linear interpolation method can be used to fill missing values ​​with the mean, median or mode; or missing values ​​can be marked as special values ​​for special processing in subsequent analysis; for erroneous data, if the format is wrong, the format is corrected; if the logic is wrong, the logic error is corrected through logical judgment, and the values ​​outside the reasonable range are corrected; for outliers, when the outliers have a large impact on the analysis results, they are directly deleted; when the outliers have a small impact on the analysis results, the outliers are replaced with values ​​within a reasonable range to obtain the data after preliminary cleaning; A dynamic time warping algorithm is used to calculate the characteristic similarity between the normal operating condition fluctuations and equipment abnormalities of the data after preliminary cleaning, and the photovoltaic panel temperature data, thermoelectric power generation unit power data, water tank temperature data, heat pipe temperature data, and thermal storage tank phase change material temperature data after precise cleaning are obtained; Based on the photovoltaic panel temperature data, heat pipe temperature data, heat storage tank phase change material temperature data, water tank temperature gradient data, and thermoelectric power generation unit power data after data cleaning, the photovoltaic panel temperature data and thermoelectric power generation unit power data are collected by a single sensor and are directly used to calculate the health index after data cleaning; the heat pipe temperature data, heat storage tank phase change material temperature data, and water tank temperature gradient data are the same parameters collected by multiple sensors, and the Kalman filter algorithm is used for data fusion to improve the accuracy and reliability of the data.

4. The condition monitoring method according to claim 1, characterized in that: The health index of the heat pipe, water tank, thermoelectric power generation unit, photovoltaic panel and heat storage tank in step S4 is calculated as follows: The heat pipe health index is calculated based on the temperature difference between the two ends of the heat pipe and the thermal conductivity efficiency. The formula is: T 冷端 and T 热端 are the temperatures of the cold and hot ends of the heat pipe, T 设定值 is the preset temperature difference; The water tank health index is calculated based on the temperature gradient and water flow rate in different areas of the water tank. The formula is: T i is the temperature of different areas in the water tank, T 平均 is the average temperature of the water tank, N is the number of sampling points; The health index of the thermoelectric power generation unit is calculated based on the power generation power and temperature difference. The formula is: P 实际 is the actual generated power, P 理论 is the theoretical power generation power; The photovoltaic panel health index is calculated based on the photovoltaic panel temperature and power generation power. The formula is: P 实际 and P 理论 are the actual power generation and theoretical power generation respectively, T 设定值 and T 实际 They are set temperature and actual temperature respectively; The heat storage tank health index is calculated based on the phase change material temperature and the temperature difference between the inside and outside. The formula is: T 内部 and T 内部 are the internal and external temperatures of the thermal storage tank, respectively.

5. The condition monitoring method according to claim 1, characterized in that: The adaptive adjustment in step S6 includes: dynamically adjusting the temperature difference at both ends of the heat pipe through a PID controller to adjust the heat pipe blockage problem; optimizing the cooling system response speed through a predictive function controller to adjust the photovoltaic panel temperature abnormality problem; and balancing the heat absorption and release rate of the phase change material through a linear quadratic regulator to adjust the heat storage tank temperature abnormality problem.

6. A status monitoring system for the photovoltaic panel waste heat coupled power generation equipment according to claim 1, characterized in that: include: Data acquisition module, used to collect real-time data on heat pipe temperature, water tank temperature gradient, thermoelectric power generation unit power, photovoltaic panel temperature, and phase change material temperature of the thermal storage tank; The data processing module includes a denoising module, a data cleaning module, and a data fusion module, where: The denoising processing module uses an adaptive filtering algorithm to denoise the photovoltaic panel temperature, heat pipe temperature and heat storage tank phase change material temperature, and uses a wavelet transform algorithm to denoise the water tank temperature gradient and the thermoelectric power generation unit power; The data cleaning module processes outliers and missing values ​​using the dynamic time warping algorithm and the isolation forest algorithm; The data fusion module uses the Kalman filter algorithm to fuse the same parameters collected by multiple sensors; The condition monitoring module defines the health index of heat pipes, water tanks, thermoelectric power generation units, photovoltaic panels, and thermal storage tanks based on data processing results, and evaluates the operating status of each component using preset thresholds; Fault diagnosis module, used to classify and locate faults based on the isolation forest model and support vector machine when the health index exceeds the threshold; The system control module is used to trigger parameter adjustment or alarm mechanism according to the fault diagnosis results.

7. The condition monitoring system according to claim 6, characterized in that: The data acquisition module includes: multiple groups of temperature sensors arranged at both ends of the heat pipe; a temperature gradient monitoring network distributed in different areas of the water tank; a K-type thermocouple module for collecting the voltage and current of the thermoelectric power generation unit; a temperature sensor set on the bottom surface of the photovoltaic panel; and an embedded sensor for monitoring the temperature difference between the inside and outside of the heat storage tank.

8. The condition monitoring system according to claim 6, characterized in that: The adaptive filtering algorithm in the denoising module is the LMS algorithm, which dynamically adjusts the filter order to match the noise spectrum; a wavelet transform algorithm is used, combined with soft threshold and hard threshold methods to perform multi-scale decomposition and reconstruction to remove noise interference in the data.

9. The condition monitoring system according to claim 6, characterized in that: The fault diagnosis module includes: using the principal component analysis algorithm to reduce the dimension of the data; performing regression analysis through the random forest algorithm to verify the data accuracy; and constructing an isolation forest model to identify abnormal trends.

10. The condition monitoring system according to claim 6, characterized in that: The system control module includes a PID controller for adjusting the temperature difference of the heat pipe, a predictive function controller for optimizing the response speed of the photovoltaic panel cooling system, and a linear quadratic regulator for balancing the heat absorption and release process of the phase change material of the heat storage tank.

Images